The Effect of Neonatal Sepsis on Risk of Autism Diagnosis

Abstract

Objective The study aimed to examine the association between neonatal sepsis and autism risk among children and whether the risk varied with the timing of exposure, child's sex, and race/ethnicity. Study Design We conducted a retrospective cohort study using electronic health records (EHR) extracted from Kaiser Permanente Southern California Health Care System. Mother–child dyads were constructed by linking records of children born to member mothers and continuing to receive care through the system during the follow-up period with those of their biological mothers (n = 469,789). Clinical health records were used to define neonatal sepsis. Diagnosis of autism was made by medical specialists. Potential confounders included maternal sociodemographic factors, obstetrical history, child's age, sex, race/ethnicity, and maternal and child medical history. Incident rates and adjusted hazard ratios (aHR) were used to estimate the associations. Results Compared with children without the diagnosis of autism, children with the condition were more likely to be from Asian/Pacific Islander descent and male sex. Exposed children showed higher rates of autism as compared with unexposed children (3.43 vs. 1.73 per 1,000 person-years, aHR: 1.67–95% confidence interval [CI]: 1.39–2.00). Both preterm (aHR: 1.47; 95% CI: 1.09–1.98) and term (aHR: 1.63; 95% CI: 1.29–2.06) births were associated with increased risk for autism. Although the magnitude of the HRs and incidence ratios for neonatal sepsis to increase autism risk varied between race ethnicities, neonatal sepsis was associated with significantly increased likelihood of autism diagnosis for all race-ethic groups except for Asian/Pacific Islanders. Although neonatal sepsis was associated with significantly increased autism risk for both boys and girls, incident rates and HR point estimates suggested that the effect may be stronger in girls. Conclusion Neonatal sepsis is associated with increased risk of autism diagnosis in preterm- and term-born children. The association was significant for both girls and boys and all race ethnicities except for Asian-Pacific Islanders. Key Points